geo-content
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python scripts (
python3 scripts/run_content.py) to perform content generation. This is the intended operation of the skill and occurs within an offline boundary as specified in the permission contract. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied content briefs and source markdown. It implements multiple safeguards, including CommonMark neutralization and HTML escaping, to mitigate the risk of processing malicious formatting or instructions embedded in source data. The workflow explicitly rejects unsupported factual claims and relies only on supplied evidence.
- [DYNAMIC_EXECUTION]: The entry point script
scripts/run_content.pydynamically appends a localsrcdirectory to the Python path (sys.path) to load thegeo_seo_hublibrary. This allows the skill to use its bundled logic while maintaining a clean structure, and the loading is limited to relative local paths within the skill's own directory structure.
Audit Metadata